Module 12 — Down the Funnel: Conversion Math

"Everybody worships the top of the funnel because it's the only number that goes up when you spend money without actually having to do anything hard. The bottom is where the truth lives, and the truth has a conversion rate that will make you want to crawl under your standing desk and stay there." — Dr. Tanya Vex, The Revenue Calculus (self-published, $1,499, available on her website between meditation retreats and LinkedIn carousels about abundance mindset)

Here's the dirty secret the demand-gen team will fight you to the goddamn death to suppress: leads are not the business. Leads are the raw material of the business, and most of them are sand — expensive, optimistically acquired, beautifully reported, totally fucking useless sand. You can pour a million leads into the top of the funnel, and if every transition between stages leaks like a knife wound in a bathtub, you'll get the same thin trickle of revenue out the bottom and a substantially fatter marketing invoice to explain to the board. The funnel is not a celebration of volume. It is a sequence of survival rates — a mortality table for deals — and your job is to know, with cold and unsentimental precision, how many die at each stage and exactly why they died there. Stop staring at the top of the funnel with stars in your eyes. The top is the easy part. Any asshole with a credit card and a LinkedIn Sales Navigator subscription can make the top of the funnel bigger. The money is in the math. The money is always in the goddamn math, and the math does not care about your beautiful MQL chart.

THE JOB

Funnel analytics is the discipline of measuring how efficiently a stranger becomes revenue, stage by stage, and locating with clinical precision where in that journey your deals are quietly hemorrhaging. It is the diagnostic layer of the entire revenue machine — the stethoscope you press to the chest of the operation to find where the heartbeat skips or stops entirely. Pipeline hygiene (Module 11) tells you whether each individual deal is real. The funnel tells you what rate of deals survive the system, in aggregate, over time. Both matter. One without the other is a half-diagnosis, which is the kind of half-assed diagnosis that gets people fired in Q4 when the miss becomes undeniable and the blame-assignment begins in earnest.

Get this right and you can do something that borders on prophecy: work backward from a revenue target and know — arithmetically, not emotionally, not on a hunch and a prayer — exactly how many leads, meetings, and opportunities you need to feed into the machine to hit The Number at the bottom. Get it wrong and you'll chase the wrong fix forever, pouring budget into the top of the funnel while the real problem bleeds out in the middle, where nobody has a dashboard with a big exciting number to defend at the all-hands because the middle doesn't get applause.

THE FUNNEL STAGES

Define these in writing, in a document everyone actually reads and agrees to, before you build a single report. The definitions are load-bearing — they are the foundation of every metric you'll ever use, and the moment someone changes them without telling anyone, everything downstream turns to shit. Change definitions silently and you corrupt every historical trend, every benchmark, and every conversation you'll have about why performance shifted:

1. Lead — a raw contact. A name and an email. Worth essentially nothing yet. Do not celebrate this stage. A mountain of leads is not a pipeline; it's a mailing list.
2. MQL (Marketing Qualified Lead) — fits the ICP and exhibited buying behavior; marketing is willing to stake a claim that it's worth a human being's time and attention.
3. SAL (Sales Accepted Lead) — sales looked at it and agreed it's worth working. The handshake. The accountability checkpoint where marketing and sales must either agree on what "qualified" means or surface their disagreement before it wastes a rep's week.
4. SQL (Sales Qualified Lead) — a rep worked it and confirmed real qualification: confirmed pain, confirmed fit, confirmed intent to evaluate. Not inferred from a web visit. Confirmed by an actual human conversation.
5. Opportunity — a qualified, dollar-valued deal in the active pipeline with exit criteria met (see Module 11). This one counts toward your pipeline number. Don't let shit in here.
6. Closed-Won — money. Actual revenue. The only stage The Number respects and the only stage that matters to the board, to your CEO, or to your continued employment.

There is also Closed-Lost — the graveyard, the data goldmine, the stage where your win-rate denominator lives and where the most important organizational intelligence gets generated and then systematically ignored by people too busy chasing the next deal to do a damn thing with the last loss. Treat it with respect. Capture reason codes. Make people look at the data. The reasons deals die here are more valuable than any pipeline number on any dashboard, and nobody wants to hear that.

THE PLAYBOOK

1. Compute the conversion rate between every stage — every goddamn one, no shortcuts

A stage conversion rate is brutally simple:

Stage conversion rate = (# that advanced to next stage) / (# that entered this stage)

Do it for every transition in the chain. Build the full table. Summarizing it into one end-to-end number and calling it a funnel analysis is the kind of thing that gets you a seat at the table for exactly one quarter before someone asks a follow-up question and realizes you've been peddling a bullshit summary where a diagnosis should live:

The end-to-end conversion (lead to won) in this example: 50 / 10,000 = 0.5%. That is normal. That is fine. That is not a crisis and you should not panic. The point is not the absolute number; the point is that you know it with precision, you know each step, and you can watch each step change over time and understand what caused the shift. A funnel you can't see is a funnel killing you in ways you haven't named yet, and it's killing you every single quarter while you argue about whether to add another field to the MQL definition.

2. Compute win rate — and for the love of god, name your goddamn denominator

Win rate = Closed-Won deals / (Closed-Won + Closed-Lost deals)

In the table above: 25% of opportunities won. Simple enough that it feels honest. The landmine: win rate from which stage? Win rate from Opportunity created versus win rate from qualified SQL are completely different numbers, and a guru will quote whichever one flatters him most depending on the composition of the room. Lance "The Closer" DiMarVo will confidently quote you a 45% win rate in a keynote without disclosing that he's measuring from first discovery call — which is not a win rate, it's a qualification filter with ambition issues.

RULE No. 19: Anyone who quotes you a win rate without naming the denominator and the stage cutoff is selling you something and hoping you don't ask. Pick the stage, write it down, enforce it consistently, and never switch denominators without announcing it clearly and updating every historical comparison. Consistency beats precision here — a consistently measured 25% win rate tells you more over two years than a "dramatically improved" 40% win rate that someone quietly redefined to exclude deals that went dark.

3. Measure stage velocity and sales-cycle length — use median or you're lying to yourself and everyone else

• Stage velocity = median days a deal spends in each stage before advancing or dying.
• Sales-cycle length = total median days from Opportunity created to Closed-Won.

Velocity is the diagnostic X-ray of your funnel. A stage with a wildly inflated dwell time is either a real buyer bottleneck — procurement grinding slowly, security review, legal doing what legal does, the whole miserable apparatus — or a process failure you own: no exit criteria, no forcing functions, nobody actually driving the deal because the rep gave up and moved on. Track it per stage. The outlier is your leak, and finding it will feel devastatingly obvious in retrospect and save you a quarter of wasted budget. And for hell's sake, use median, not mean — one 400-day enterprise whale will drag your arithmetic mean into a fantasy that makes every rep's cycle look broken, and you'll spend a whole quarter trying to "fix" a problem that's actually just one account with a dysfunctional procurement process and a general counsel who treats every vendor like a hostile witness. Median tells you what's typical. Mean tells you about the outliers. You need to know typical first, or you'll be chasing ghosts.

4. Find the leak — diagnose where deals die before you spend a dollar on the top

The leaky funnel is the whole game. This is the method. Follow it in order:

1. Build the stage-by-stage conversion table above for your current measurement period.
2. Benchmark every transition against your own trailing average — your own history, your own data, not whatever Lance DiMarVo scraped from a survey of 200 companies with no disclosed methodology. Your funnel is yours. Own it.
3. Find the transition with the worst conversion relative to its own historical norm. That transition is your leak. That is where to focus before you approve one more dollar of top-of-funnel spend.
4. Diagnose by location:
   • Leak at the top (Lead → MQL) → targeting or lead-quality problem. You are buying sand and paying content marketing rates for it. The ICP definition is wrong, the list is garbage, or the lead-scoring model has been flattering noise for months.
   • Leak in the middle (MQL → SQL, SAL → SQL) → qualification or speed-to-lead problem (see Module 10). Marketing and sales do not agree on what "qualified" means and the disagreement is costing you real deals that die from neglect before anyone gets to them.
   • Leak at the bottom (Opp → Won) → sales execution, pricing, or competitive problem. These deals are real and you're losing them in the room, on the proposal, in the negotiation, or to a competitor with a better offering. This is the most expensive leak because you've burned the most money and time to reach it.

The instinct — the almost universal, extremely expensive, reliably wrong instinct — is to fix the top because it's visible, it's exciting, it has a budget owner with a Patagonia vest and strong opinions about attribution. The most expensive leak is almost always in the middle, where sales-touched, hard-won, expensive-to-generate MQL-to-SQL deals evaporate quietly without anyone owning the failure or logging the loss with any damn context. Nobody raises their hand for the middle. Fix it anyway. Fix it first. The top of the funnel can wait.

5. Do the backward pipeline math from the revenue target — the most useful calculation in the whole building

Work upward from The Number. Not down from what you have — up from what you need. This is how you find out in June whether your annual plan is grounded in arithmetic or in some combination of cocaine optimism and wishful thinking approved by a board that didn't ask for the supporting math:

Pipeline needed ($) = Revenue target / Win rate
Opportunities needed = Pipeline needed / Average deal size (ACV)
SQLs needed = Opportunities needed / (SQL → Opp conversion rate)
SALs needed = SQLs needed / (SAL → SQL conversion rate)
MQLs needed = SALs needed / (MQL → SAL conversion rate)
Leads needed = MQLs needed / (Lead → MQL conversion rate)

Worked example. Target = $5,000,000. Win rate = 25%. ACV = $50,000.

• Pipeline needed = $5,000,000 / 0.25 = $20,000,000 in qualified opportunity value. This is your 3x coverage requirement — and notice that the inverse of a 25% win rate is mathematically 4x, which means everyone who tells you "3x is enough" is already building in a planned miss and hoping the quarter bails them out. 3x pipeline at a 25% win rate produces $3.75M. You need $5M. Do the fucking math before you accept the plan.
• Opportunities needed = $20,000,000 / $50,000 = 400 opportunities.
• At 50% SQL → Opp: 800 SQLs.
• At 67% SAL → SQL: ~1,195 SALs.
• At 50% MQL → SAL: ~2,390 MQLs.
• At 12% Lead → MQL: ~19,900 leads.

Now you know the truth and you can't unknow it: hitting $5M requires roughly 20,000 leads through this particular funnel. If marketing can generate 8,000, your annual plan is a hallucination. You found out now — in planning, in June — instead of in December when the miss is real and everyone's looking for someone to blame and the RevOps analyst who ran this math in a slide is mysteriously unavailable.

6. Compute pipeline velocity — the master equation, all four levers

This is the formula that should be tattooed on the inside of every RevOps person's eyelids:

                    # of Opportunities  ×  Win Rate  ×  Average Deal Value (ACV)
Pipeline Velocity = ─────────────────────────────────────────────────────────────
                                    Sales-Cycle Length (days)

Example: 400 opps × 25% × $50,000 / 90 days = $5,000,000 / 90 = ~$55,556 of revenue per day.

The beauty of this equation is not the number it produces — it's what it forces you to see. It shows you the four and only four levers that move revenue. To grow the machine, you either:

(a) Create more opportunities — top-of-funnel volume, better ICP targeting, more pipeline generation activity
(b) Win a higher percentage — sales execution, competitive differentiation, better discovery, methodology enforcement
(c) Raise ACV — bigger deals, better packaging, move upmarket, stop discounting your face off in a panic at the end of every goddamn quarter
(d) Shorten the cycle — speed up the buyer's process, remove internal friction, stop letting deals rot in Stage 3 for three months because nobody's driving them

That is the complete and exhaustive list. Every initiative anyone proposes — every tool, every play, every reorg, every consultant with a Moleskine and strong opinions, every enablement program, every methodology course Lance DiMarVo is hawking at $2,000 a seat with a money-back guarantee that requires seven forms — must demonstrably move one of those four numbers or it is theater. Make every person in the room point at the lever before you approve a cent of the spend. If they can't name the lever, they don't actually understand what they're proposing and they're betting that sunk-cost fallacy and meeting fatigue carry them through to approval. This single question — which lever? — will save you from a staggering amount of expensive, well-intentioned, deeply confident bullshit.

7. Separate vanity volume from real conversion — or stop pretending you're doing analytics

A bigger top of funnel that converts worse is not growth. It is a more expensive way to produce the same revenue, with a higher CAC, more burned rep hours, and a better-looking chart that hides everything that matters. Always — always — pair every volume metric with its conversion rate. If MQLs doubled but MQL → SQL halved, you didn't generate more demand. You bought noise, named it, handed it to your reps, and made them chase it on their precious quota time. That is not a success story. That is a waste of money masquerading as a success story because the top number went up and the board said "great job."

RULE No. 20: Volume without conversion is just CAC with better marketing. Pair every count with its rate or admit you're hiding something.

HOW IT GOES TO HELL

The VP of Vibes opens the board meeting with "MQLs are up 60% quarter over quarter!" and receives actual applause from a room of sophisticated investors who should absolutely know better. Nobody asks what happened to MQL → SQL conversion, which crashed from 50% to 28%, because the new MQLs are from a list he bought at a conference booth to make the top number move before his performance review. The funnel got wider and shallower — more leads, worse leads, same revenue, substantially higher cost — and he stood up in front of the board and called it momentum with a straight face. Vanity volume is the oldest con in B2B SaaS and it works every single goddamn time, because the top number is the easiest one to inflate and the only one that instinctively, emotionally feels like winning. The board applauds. The RevOps analyst stares at the floor and wonders what to put on her résumé.

The Guru — Lance "The Closer" DiMarVo, whose LinkedIn banner says "PIPELINE IS A MINDSET" in a font that communicates aggressive mediocrity — sells a course built on a single end-to-end conversion rate scraped from a survey of 200 companies in three different industries with no disclosed methodology, applied to your specific business as though it were gospel. Your funnel is not his funnel. Your ACV is not his ACV. Your ICP is not his demographic. Industry benchmarks are weather reports from other planets: interesting as reference, dangerously misleading as targets. Use your own trailing data or use nothing. Your own history is the only honest benchmark you've got and it's sitting right there in your CRM, which you pay for every month and consult less often than you should.

The RevOps Martyr built a flawless funnel dashboard with drill-downs by segment, by rep, by channel, by cohort — genuinely beautiful, genuinely useful work. Then someone — a sales leader, a marketing director, an excited new SDR manager who didn't read any documentation — quietly redefined "MQL" in March without telling anyone, so the historical trend is now comparing two completely different populations that share a label and nothing else. Every conversion rate silently shifted. She spent two months chasing a leak that was actually a definitional change. A funnel metric is only as good as the stability of its stage definitions, which must be written down, versioned, and changed only through a formal process that produces a timestamp, a note in the data dictionary, and a communication to every person who uses the metrics. Change a definition silently, break the trend. Break the trend, lose the diagnostic. Lose the diagnostic, and you're back to managing by gut feel, which is what everyone was doing before she built the whole goddamn thing.

FIELD RULES

• RULE No. 12a: The top of the funnel is the easiest number to fake and the least important one to optimize. Watch the transitions — that's where the truth about your revenue machine actually lives.
• RULE No. 12b: A conversion rate without a named denominator, a stable stage definition, and a consistent measurement window is a rumor dressed up in a spreadsheet.
• RULE No. 12c: Pipeline coverage = target ÷ win rate. At 25% win rate, you need 4x pipeline, not the comfortable 3x everyone keeps quoting. Do the arithmetic before you approve the plan.
• RULE No. 12d: Volume without conversion is just CAC with better marketing. Pair every count with its rate or stop calling it a metric.
• RULE No. 12e: Pipeline velocity gives you exactly four levers — more opps, higher win rate, bigger ACV, shorter cycle. Every proposed initiative must name its lever or it doesn't get funded. No exceptions.
• RULE No. 12f: Benchmark against your own history. Other companies' conversion rates are weather from other planets — interesting, occasionally instructive, irrelevant as a personal target.

---

From the field: I once watched a company set a $12M annual plan on a funnel that, when I ran the backward math against their own stated conversion rates, required more qualified discovery meetings than there were target accounts in their entire addressable market. The arithmetic was disqualifying before the planning deck was even finished. I showed them the math in a slide. The CEO called it "conservative framing." The CMO called it "not the spirit of the exercise." They funded it anyway, missed by half, and fired the RevOps analyst who had put the math in writing and asked everyone to look at it before committing. The math was never wrong. The math is never fucking wrong. People just find it unbearable to confront it sober. Expense line: one calculator that still works, one drink that helped, and the cold, quiet satisfaction of having said so in writing before it happened, dated and timestamped, permanently in the record.